Previous experiments indicate that glucose turnover measured with [6-3H] glucose during hyperinsulinemia underestimated glucose turnover measured with [6-14]C glucose. Experiments in the last year have indicated that this error is due to the presence of a nonglucose contaminant in the [6-3H] infusate. The contaminant can now account for discrepancies in insulin action in man.